On Thu, Mar 02, 2000 at 09:12:07AM +0100, Anders Holm wrote: > > I saw Your first few comments, thank you. > > - Item ids: Using unique (within a table) ids meaning absolutely is > the standard way to have guarantied unique references as time > devellops. But in many cases names and other values are known to be > unique. > MySQL can enforce uniqueness of data fields. This means that uniqueness of "id"s can be guarantied for any kind of "id" (artificial or meaningful). > > - Time stamps: as ids is dangerous. Time stamps change (or may change) > when items are updated. The standard time stamps I used in MySql are only > with one second precission and thus not guarantied uniqe. > Again, if the uniqueness is enforced by the database, this is not a problem. I do not think we should use the MySQL "TIMESTAMP" data type, but use an "INT" data type filled by the "unix time" generated by the application that puts the data into database. Is there still a problem with uniqueness? > > Perl: I have not used Perl with data bases. But it seems that whatever is > simple in SQL terms is doable (and equally simple?) in Perl. > If you want to query the database through the web, perl and CGI.pm are your men. -- Konstantin Olchanski Physics Department, Brookhaven National Laboratory, Long Island, New York olchansk@bnl.gov
This archive was generated by hypermail 2b29 : Thu Mar 02 2000 - 14:48:31 EST